| [925] | 1 | #include "mex.h" |
|---|
| 2 | |
|---|
| 3 | /* |
|---|
| 4 | * $Id: xml_findstr.c 6480 2015-06-13 01:08:30Z guillaume $ |
|---|
| 5 | * Guillaume Flandin <guillaume@artefact.tk> |
|---|
| 6 | */ |
|---|
| 7 | |
|---|
| 8 | /* |
|---|
| 9 | Differences with built-in findstr: |
|---|
| 10 | - allows to search only the n first occurences of a pattern |
|---|
| 11 | - allows to search only in a substring (given an index of the beginning) |
|---|
| 12 | |
|---|
| 13 | MATLAB hack: |
|---|
| 14 | - doesn't use mxGetString to prevent a copy of the string. |
|---|
| 15 | - assumes MATLAB stores strings as unsigned short (Unicode 16 bits) |
|---|
| 16 | matrix.h: typedef uint16_T mxChar; |
|---|
| 17 | (that's the case for MATLAB 5.*, 6.* and 7.* but MATLAB 4.* stores |
|---|
| 18 | strings as double and GNU Octave as char, see src/mxarray.h) |
|---|
| 19 | */ |
|---|
| 20 | |
|---|
| 21 | /* Comment the following line to use standard mxGetString (slower) */ |
|---|
| 22 | #if !defined (HAVE_OCTAVE) |
|---|
| 23 | #define __HACK_MXCHAR__ |
|---|
| 24 | #endif |
|---|
| 25 | |
|---|
| 26 | void mexFunction(int nlhs, mxArray *plhs[], int nrhs, const mxArray *prhs[]) { |
|---|
| 27 | |
|---|
| 28 | unsigned int i, j, stext, spattern, nbmatch = 0, ind = 1, occur = 0, nboccur = 0; |
|---|
| 29 | #ifdef __HACK_MXCHAR__ |
|---|
| 30 | unsigned short int *text = NULL, *pattern = NULL; |
|---|
| 31 | #else |
|---|
| 32 | char *text = NULL, *pattern = NULL; |
|---|
| 33 | #endif |
|---|
| 34 | unsigned int *k = NULL; |
|---|
| 35 | mxArray *out = NULL; |
|---|
| 36 | |
|---|
| 37 | /* Check for proper number of arguments. */ |
|---|
| 38 | if ((nrhs == 0) || (nrhs == 1)) |
|---|
| 39 | mexErrMsgTxt("Not enough input arguments."); |
|---|
| 40 | else if (nrhs > 4) |
|---|
| 41 | mexErrMsgTxt("Too many input arguments."); |
|---|
| 42 | else if (nlhs > 1) |
|---|
| 43 | mexErrMsgTxt("Too many output arguments."); |
|---|
| 44 | |
|---|
| 45 | /* The input TEXT must be a string */ |
|---|
| 46 | if (!mxIsChar(prhs[0])) |
|---|
| 47 | mexErrMsgTxt("Inputs must be character arrays."); |
|---|
| 48 | stext = mxGetM(prhs[0]) * mxGetN(prhs[0]); |
|---|
| 49 | #ifdef __HACK_MXCHAR__ |
|---|
| 50 | text = mxGetData(prhs[0]); |
|---|
| 51 | #else |
|---|
| 52 | text = mxCalloc(stext+1, sizeof(char)); |
|---|
| 53 | mxGetString(prhs[0], text, stext+1); |
|---|
| 54 | #endif |
|---|
| 55 | |
|---|
| 56 | /* The input PATTERN must be a string */ |
|---|
| 57 | if (!mxIsChar(prhs[1])) |
|---|
| 58 | mexErrMsgTxt("Inputs must be character arrays."); |
|---|
| 59 | spattern = mxGetM(prhs[1]) * mxGetN(prhs[1]); |
|---|
| 60 | #ifdef __HACK_MXCHAR__ |
|---|
| 61 | pattern = mxGetData(prhs[1]); |
|---|
| 62 | #else |
|---|
| 63 | pattern = mxCalloc(spattern+1, sizeof(char)); |
|---|
| 64 | mxGetString(prhs[1], pattern, spattern+1); |
|---|
| 65 | #endif |
|---|
| 66 | |
|---|
| 67 | /* The input INDEX must be an integer */ |
|---|
| 68 | if (nrhs > 2) { |
|---|
| 69 | if ((!mxIsNumeric(prhs[2]) || (mxGetM(prhs[2]) * mxGetN(prhs[2]) != 1))) |
|---|
| 70 | mexErrMsgTxt("Index input must be an integer."); |
|---|
| 71 | ind = (unsigned int)mxGetScalar(prhs[2]); |
|---|
| 72 | if (ind < 1) |
|---|
| 73 | mexErrMsgTxt("Index must be greater than 1."); |
|---|
| 74 | } |
|---|
| 75 | |
|---|
| 76 | /* The input OCCUR must be an integer */ |
|---|
| 77 | if (nrhs == 4) { |
|---|
| 78 | if ((!mxIsNumeric(prhs[3]) || (mxGetM(prhs[3]) * mxGetN(prhs[3]) != 1))) |
|---|
| 79 | mexErrMsgTxt("Index input must be an integer."); |
|---|
| 80 | nboccur = (unsigned int)mxGetScalar(prhs[3]); |
|---|
| 81 | } |
|---|
| 82 | |
|---|
| 83 | /* Find pattern in text */ |
|---|
| 84 | for (i=ind-1;i<stext;i++) { |
|---|
| 85 | for (j=0;j<spattern && i+j<stext;j++) { |
|---|
| 86 | if (pattern[j] == text[i+j]) { |
|---|
| 87 | if (j == spattern-1) { |
|---|
| 88 | nbmatch += 1; |
|---|
| 89 | k = mxRealloc(k,nbmatch*sizeof(unsigned int)); |
|---|
| 90 | k[nbmatch-1] = i+1; |
|---|
| 91 | if (++occur == nboccur) i = stext; |
|---|
| 92 | } |
|---|
| 93 | } |
|---|
| 94 | else break; |
|---|
| 95 | } |
|---|
| 96 | } |
|---|
| 97 | |
|---|
| 98 | /* Allocate output */ |
|---|
| 99 | out = mxCreateDoubleMatrix((nbmatch) ? 1:0, nbmatch, mxREAL); |
|---|
| 100 | |
|---|
| 101 | /* Copy index array into output */ |
|---|
| 102 | for (i=0;i<nbmatch;i++) |
|---|
| 103 | mxGetPr(out)[i] = (double)k[i]; |
|---|
| 104 | |
|---|
| 105 | /* Assign pointer to output */ |
|---|
| 106 | plhs[0] = out; |
|---|
| 107 | |
|---|
| 108 | /* Free memory */ |
|---|
| 109 | if (k) mxFree(k); |
|---|
| 110 | } |
|---|
